Watchdog problem

[Fenice]

wdcollect[327]: Statistics stack owerflow. Stack contains now 1000 unsaved entries, some of them will be removed.

This entry is repeated continuously in message log.

How can I fix it?

 

[luribe]

I'm having the same problem, so i have to deactivate watchdog.

 

[bigtank]

Hi @ALL

I made a call for this problem...

Quotation from an engineer:

wdcollect keeps some time statistic information in his stack, then wdcollect try to save this data in mysql db. After data saving in mysql db this dwcollect stack is cleared.

This message "wdcollect[13359]: Statistics stack owerflow. Stack contains now 1000 unsaved entries, some of them will be removed" occurs if wdcollect couldn't connect to db and it try to connect to db again after unsuccessful connection.

To solve this problem I killed the running wdcollect process and init started wdcollect himself:


# tail -f /usr/local/psa/var/modules/watchdog/log/wdcollect.log
7200 Apr 5 09:54:34 [info] Service is being stopped
7200 Apr 5 09:54:34 [info] There are unsaved messages in the database messages queue. Trying to save the messages.
7200 Apr 5 09:54:34 [warning] There are unsaved messages in the database messages queue. All the unsaved messages will be discarded.
7200 Apr 5 09:54:34 [info] Server stopped
7200 Apr 5 09:54:34 [info] Language en-US is used for sending e-mail messages.
7200 Apr 5 09:54:34 [info] Database server connection has been established.
7200 Apr 5 09:54:34 [info] Server started.

But all statistics data from last days are lost.

I forwarded this issue to our developers to investigation.

I hope this helps everybody with the same problem

bigbtank

 

[kerpau]

Just worked through an issue like this one. It turns out that there was an additional (10.X.X.X) IP address configured in Plesk that didn't exist on the network interface. Seems that watchdog was monitoring this IP address, erroring out and hanging up the server. I removed the IP from Plesk, restarted Plesk services and it seems to have resolved the issue. Hope this info helps someone else out!

 

[bonesadams]

Hi,

I had the same issue before with 8.0. Found that some hosting companies are doing image of Plesk on Windows/*nix OS and then just put it on the all servers they provide.
It was funny that there was appearing an IP which was not configured, but a lot of the services are configured to use it.
A simple fix will be to use this command :

/usr/local/psa/bin/ipmanage --remove x.x.x.x

Some of the services will be automatically fixed, but if not, then you have to reinstall the service via ssh

 

[-yom-]

thanks a lot !

thanks for that solution
it helped me to solve that issue

i'm running plesk 8.0.1 on FC4, and just needed to kill the "wdcollect" process to solve the problem
it restarted automatically

tail -f /usr/local/psa/var/modules/watchdog/log/wdcollect.log returns :

3600 Dec 12 09:35:40 [info] Service is being stopped
3600 Dec 12 09:35:40 [info] Server stopped
3600 Dec 12 09:35:40 [info] Language fr-FR is used for sending e-mail messages.
3600 Dec 12 09:35:40 [info] Database server connection has been established.
3600 Dec 12 09:35:40 [info] Server started.

everything's fine, watchdog show statistics again!

 

[keff]

In my plesk installation watchdog still doesn't shows the statistics. In wdcollect.log after restarting wdcollect proces I have:
wdcollect: Statistics stack owerflow. Stack contains now 1001 unsaved entries, some of them will be removed.
wdcollect: Statistics stack owerflow. Stack contains now 1001 unsaved entries, some of them will be removed.
wdcollect: Statistics stack owerflow. Stack contains now 1001 unsaved entries, some of them will be removed.

What else can I do??

 

[IckZ]

same problem here, what can I do? I restarted the Service and looked after some other things, but no solution.

thanks for any answers...

 

[-yom-]

it happened again

have had this problem on another webserver
Plesk 8.2.1 on FC7

same problem, same issue, same solution

thanks again !!!!